Why the Crash Arrives Later

The subject now is the delayed macroeconomic response to credit-supply financial liberalization: why output can improve first, why crisis risk rises later, and what the historical evidence can actually establish. The goal is to read the estimates as a timeline and a distribution of outcomes rather than as a single verdict on deregulation. That distinction makes the evidence more useful for judging a reform whose initial results are visible long before its possible costs.

The hump-shaped result

The Federal Reserve study finds a distinct time profile after reforms that directly loosen the supply of bank credit. Real GDP per person rises in the first few years, increasing by about 0.74 percent cumulatively over the first three years in the reported local-projection estimates. The response peaks around the fourth year and then fades. By roughly the eighth year, cumulative GDP per person has returned to its pre-liberalization trend. Annual growth becomes negative in the medium term rather than continuing the initial acceleration.

The shape matters more than the first positive number. A level can rise and later return to trend even if the economy has enjoyed several years of higher activity. The finding does not say that every country ends exactly where it began, or that households receive no lasting benefit from newly available finance. It says that the average post-reform output path in this sample does not preserve the initial boost once the later slowdown is included.

The credit response helps explain the timing. A direct supply reform can relax a financing constraint immediately, but productive investment takes time to plan and complete. Property and construction activity can respond more quickly to easier borrowing, while the liabilities supporting that activity accumulate over several years. The initial credit expansion can therefore lift measured activity before the stock of leverage becomes large enough to affect repayment, funding, and aggregate demand. The lag is part of the economic process, not a statistical nuisance to be averaged away.

The sectoral allocation makes the reversal more understandable. New credit flows disproportionately toward non-tradable activities, particularly construction and real estate. Those sectors can generate employment, transactions, and visible investment during the boom. They are also often more dependent on collateral values and domestic financing conditions. When property prices stop rising or borrowers have to reduce debt, the sector that responded strongly to new credit can become a source of contraction.

Crisis risk follows a still later path. The study finds little short-run change in the likelihood of a systemic banking crisis immediately after deregulation. From about the three-year horizon onward, the estimated risk rises. For a one-unit increase in the credit-supply index, the reported marginal effect is about 2.8 percentage points at three years and about 9.8 percentage points at eight years, relative to an economy whose regulatory framework did not change. The eight-year estimate is large compared with the sample’s unconditional crisis probability of 1.9 percent.

These are not probabilities attached to a particular country’s future reform. They are marginal-effect estimates tied to the index, the sample, the crisis chronology, the controls, and the chosen horizon.